High speed Inductive proximity sensor

Discussion in 'The Projects Forum' started by blackflag267x, Sep 10, 2014.

  1. blackflag267x

    Thread Starter New Member

    Sep 10, 2014
    6
    0
    Looking to build (or buy) an inductive proximity sensor that can detect copper bullets moving at around 900 m/s. My experience with circuits is small so forgive any misunderstandings in the following. I am always open to learning opportunities.

    -Sensor composed of a wire coil, oscillator, Schmitt trigger (or just comparator?), and (maybe?) a demodulator. The signal from the sensor will be read by a microprocessor (right now I have an arduino uno)

    -oscillator induces a current in the coil creating a magnetic field. When the copper passes through this field, the field changes, inducing a change in the current or affecting the oscillator in some way (need pointers on how this happens). Because the bullet is moving quickly, the sensor will have to react quickly. I believe this means the coil and oscillator needs low reactance and thus low inductance.

    -the schmitt trigger (i think this is necessary over just a comparator because it can reduce the noise of the signal) will need to be able to detect small changes in the current and thus voltage. Not sure what magnitude but probably sub-microvolt differences. It also needs to have a fast slew rate (I think this is the term?) because the bullet travels through one centimeter in about 11 microseconds.

    -not sure if the demodulator is necessary to convert information for the microprocessor to read or if the schmitt trigger can do this.

    My two challenges at this point (other than overall understanding) are determining how the coil should be built and oriented and what device (both model and type) can be used to detect current changes.

    TL;DR: want to detect quickly moving (900 m/s) copper with a magnetic field and some sort of amplifier (schmitt trigger?), need pointers on inductance in coils and detection devices

    Thanks for any pointers!
     
  2. wayneh

    Expert

    Sep 9, 2010
    12,087
    3,027
    Do you have control over mechanics? I mean, what is the range of clearances from the moving object to the surface of the coil?

    Is there a reason you have dismissed optical methods?
     
  3. crutschow

    Expert

    Mar 14, 2008
    12,977
    3,221
    What level of accuracy and precision are you wanting.
     
  4. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,127
    266
    I assume you intend to have two measuring coils and measure the speed of the bullet by the timing between the two coils?
    So your detector must have a very consistent rise time?

    Hmmm? seems like you may be able to accomplish this with a DC bias on the coil that creates a strong magnetic field, the copper slug passing through will experience eddy currents that will disturb the bias on the coil, should be easy to detect with and AC coupled amp and comparator.

    Here is a link to a commercial product- probably uses permanent magnets and coils.

    http://www.realguns.com/articles/360.htm
     
    Last edited: Sep 11, 2014
  5. blackflag267x

    Thread Starter New Member

    Sep 10, 2014
    6
    0
    The sensor could be within 0.5" of the bullet. For my purposes specifically, I thought it might be simpler to build an electromagnetic sensor than an optical one because optical sensors are so sensitive to variations in ambient light, lighting conditions, lighting changes, etc. I figured the measured change in signal from both sensors from a detection event would be pretty similar, but the electromagnetic sensor would have less noise and fewer false readings to factor out.

    Within 5 m/s of true value +/-2 or 3 m/s would be ideal, slightly less precision would be ok. if that makes sense

    Yes, two coils. The rise time would need to be in the tens of microseconds for optimal detection.
    That product is basically what I'm trying to emulate. I want to build it because that's much more satisfying (ultimately) than just spending the dough.
    I was originally thinking I'd go with permanent magnets and coils but thought that the field from just an ac current in a coil would be easier to disturb and thus detect. What makes you think permanent magnets are used? Would that make detection easier or more difficult?
    I understnad the DC bias and eddy currents; what's an AC coupled amp? The internet is not giving a great explanation. I assume it amplifies the signal that the comparator then detects, but what is the AC coupled part?
     
  6. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,127
    266
    Google "variable reluctance sensor" the definition is not strictly correct in this case because your target is non ferrous.
    If you use an AC excitation for your coil, you will need to demodulate the signal as a phase or frequency change, this requires analysis of the signal over time, which creates delays and timing jitter. Detecting the bullet with a constant bias magnetic field is much simpler, you just need an amplifier that can react to the fast pulse created in the coil. This is where the AC coupling comes in, you only want to amplify the AC signal, not any bias or offsets. Permanent magnets require no power, you could use a plain coil with a bias current, but it's going to eat up lots of power.
     
  7. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    I, too, like to do things a different way, but having spent a lot of time shooting with chronographs, I think that placing sensors that close to the bullet path is problematic. It's amazing how many optical sensors get shot up, and the bullet path doesn't have to be anywhere that close. In addition, the placement of small translucent panel above the optical sensors all but eliminates problems with detection. But, I understand wanting to "try it anyway," so I wish you the best of luck.
     
  8. alfacliff

    Well-Known Member

    Dec 13, 2013
    2,449
    428
    a coil the bullet travels through, with a magnet to bias the coil, a conductive bullet traveling through the coi8l will induce a pulse that could be used to start a counter, another coil a distance away (about 1 foot with a 100 mhz clock for the counter) to stop the count. with a higher frequency clock, the coils could be closer together, possible on a common core, possible ceramic fastened to the end of the barre.
    I thought this one up a few years ago.
     
  9. blackflag267x

    Thread Starter New Member

    Sep 10, 2014
    6
    0
    sounds like you have more experience than I do, so I'm just voicing my thoughts here: with an optical sensor setup the sensor is placed slightly downrange, increasing the chance of hitting it with some sighting errors; with the commercial electromagnetic setup, and how I'm envisioning my attempt, the device is actually attached to the barrel, drastically reducing the chance of hitting the setup. If I can't get the magnets to work though, I'll switch to optical. Also, I like the tardigrade.
     
  10. blackflag267x

    Thread Starter New Member

    Sep 10, 2014
    6
    0
    How do you create a constant bias magnetic field with a permanent magnet? Or is that what a permanent magnet's field is? And then where does the AC signal that would be amplified come from - is that the pulse created by the disturbance from the bullet?

    Secondarily, would the small amount of copper disturb the permanent magnet's strong field enough to be detectable? This is why I shifted to the AC driven electromagnet - a weaker field that is more easily disturbed (but that's just an assumption). I guess I need to do more experimenting.
     
  11. alfacliff

    Well-Known Member

    Dec 13, 2013
    2,449
    428
    as long as the bullet travels through the magnetic field and the coil is in the field too, any conductive material traveling through the field will cause a change in the field which will induce a pulose in the coil.
     
  12. tracecom

    AAC Fanatic!

    Apr 16, 2010
    3,869
    1,393
    Sounds plausible. I'll be watching the thread to see your progress.
     
  13. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,127
    266
    Put the sensor ( a strong permanent magnet with a coil around it ) perpendicular to the bullet path, so the bullet crosses close to one end of the magnet.
    A permanent magnet will create a very strong permanent bias field for the bullet to interact with.
    The perpendicular orientation allows the bullet's copper jacket to cut the looping magnetic field lines emerging from the end of the magnet, which will create the eddy currents that we desire. Some VR sensors have a ferromagnetic 'cup' that provides a flux path around the outside of the coil.

    Trying to shoot THROUGH the coil will be very problematic and would require a power hungry bias current in the coil.

    You could just BUY a VR sensor off the shelf. images (1).jpg
     
  14. blackflag267x

    Thread Starter New Member

    Sep 10, 2014
    6
    0
    Ok, great. That was my original approach. I'll probably try to make the sensor setup (loop and magnet) before buying one. Now I need to figure out what signal detection setup is appropriate for the speed and level of detection I need. This was the first thing I came across:

    http://datasheets.maximintegrated.com/en/ds/MAX9924-MAX9927.pdf

    Some simple specs/my speculations:
    4.5 - 5.5 V supply voltage (would work with arduino uno)
    down to 50 nS propagation delay (way quicker than I'd need)
    input bias and offset currents on the nA level (probably suitable for detection)
    Best of all they provide a free sample of a dual input sensor

    There are some things on that datasheet I don't have time to understand until the weekend, so I may be jumping the gun on that specific part. But hey, if its free...
     
  15. blackflag267x

    Thread Starter New Member

    Sep 10, 2014
    6
    0
    Nevermind, not free
     
Loading...